codersan/validadted_e5SmallFa_onV9f

SentenceTransformer based on intfloat/multilingual-e5-small

This is a sentence-transformers model finetuned from intfloat/multilingual-e5-small. It maps sentences & paragraphs to a 384-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.

Model Details

Model Description

• Model Type: Sentence Transformer
• Base model: intfloat/multilingual-e5-small
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 384 dimensions
• Similarity Function: Cosine Similarity

Model Sources

• Documentation: Sentence Transformers Documentation
• Repository: Sentence Transformers on GitHub
• Hugging Face: Sentence Transformers on Hugging Face

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: BertModel 
  (1): Pooling({'word_embedding_dimension': 384, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
  (2): Normalize()
)

Usage

Direct Usage (Sentence Transformers)

First install the Sentence Transformers library:

pip install -U sentence-transformers

Then you can load this model and run inference.

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("codersan/validadted_e5SmallFa_onV9f")
# Run inference
sentences = [
    'برای تبدیل شدن به نویسنده برتر Quora ، چند بازدید و پاسخ لازم است؟',
    'چگونه می توانم نویسنده برتر Quora شوم ، از صعود بیشتر و آمار بهتر استفاده کنم؟',
    'من به دنبال خرید دوچرخه جدید هستم.Suzuki Gixxer 155 یا Honda Hornet 160r.کدام یک را بخرید؟',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 384]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]

Training Details

Training Dataset

Unnamed Dataset

• Size: 131,157 training samples
• Columns: anchor and positive
• Approximate statistics based on the first 1000 samples:

  	anchor	positive
  type	string	string
  details	min: 6 tokens mean: 16.81 tokens max: 97 tokens	min: 5 tokens mean: 16.59 tokens max: 59 tokens

• Samples:

  anchor	positive
  وقتی سوال من به عنوان "این سوال ممکن است به ویرایش نیاز داشته باشد" چه کاری باید انجام دهم ، اما نمی توانم دلیل آن را پیدا کنم؟	چرا سوال من به عنوان نیاز به پیشرفت مشخص شده است؟
  چگونه می توانید یک فایل رمزگذاری شده را با دانستن اینکه این یک فایل تصویری است بدون دانستن گسترش پرونده یا کلید ، رمزگشایی کنید؟	چگونه می توانید یک فایل رمزگذاری شده را رمزگشایی کنید و بدانید که این یک فایل تصویری است بدون اینکه از پسوند پرونده اطلاع داشته باشید؟
  احساس می کنم خودکشی می کنم ، چگونه باید با آن برخورد کنم؟	احساس می کنم خودکشی می کنم.چه کاری باید انجام دهم؟

• Loss: MultipleNegativesRankingLoss with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim"
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 12
• learning_rate: 5e-06
• weight_decay: 0.01
• num_train_epochs: 1
• warmup_ratio: 0.1
• push_to_hub: True
• hub_model_id: codersan/validadted_e5SmallFa_onV9f
• eval_on_start: True
• batch_sampler: no_duplicates

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• prediction_loss_only: True
• per_device_train_batch_size: 12
• per_device_eval_batch_size: 8
• per_gpu_train_batch_size: None
• per_gpu_eval_batch_size: None
• gradient_accumulation_steps: 1
• eval_accumulation_steps: None
• torch_empty_cache_steps: None
• learning_rate: 5e-06
• weight_decay: 0.01
• adam_beta1: 0.9
• adam_beta2: 0.999
• adam_epsilon: 1e-08
• max_grad_norm: 1
• num_train_epochs: 1
• max_steps: -1
• lr_scheduler_type: linear
• lr_scheduler_kwargs: {}
• warmup_ratio: 0.1
• warmup_steps: 0
• log_level: passive
• log_level_replica: warning
• log_on_each_node: True
• logging_nan_inf_filter: True
• save_safetensors: True
• save_on_each_node: False
• save_only_model: False
• restore_callback_states_from_checkpoint: False
• no_cuda: False
• use_cpu: False
• use_mps_device: False
• seed: 42
• data_seed: None
• jit_mode_eval: False
• use_ipex: False
• bf16: False
• fp16: False
• fp16_opt_level: O1
• half_precision_backend: auto
• bf16_full_eval: False
• fp16_full_eval: False
• tf32: None
• local_rank: 0
• ddp_backend: None
• tpu_num_cores: None
• tpu_metrics_debug: False
• debug: []
• dataloader_drop_last: False
• dataloader_num_workers: 0
• dataloader_prefetch_factor: None
• past_index: -1
• disable_tqdm: False
• remove_unused_columns: True
• label_names: None
• load_best_model_at_end: False
• ignore_data_skip: False
• fsdp: []
• fsdp_min_num_params: 0
• fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
• fsdp_transformer_layer_cls_to_wrap: None
• accelerator_config: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
• deepspeed: None
• label_smoothing_factor: 0.0
• optim: adamw_torch
• optim_args: None
• adafactor: False
• group_by_length: False
• length_column_name: length
• ddp_find_unused_parameters: None
• ddp_bucket_cap_mb: None
• ddp_broadcast_buffers: False
• dataloader_pin_memory: True
• dataloader_persistent_workers: False
• skip_memory_metrics: True
• use_legacy_prediction_loop: False
• push_to_hub: True
• resume_from_checkpoint: None
• hub_model_id: codersan/validadted_e5SmallFa_onV9f
• hub_strategy: every_save
• hub_private_repo: None
• hub_always_push: False
• gradient_checkpointing: False
• gradient_checkpointing_kwargs: None
• include_inputs_for_metrics: False
• include_for_metrics: []
• eval_do_concat_batches: True
• fp16_backend: auto
• push_to_hub_model_id: None
• push_to_hub_organization: None
• mp_parameters:
• auto_find_batch_size: False
• full_determinism: False
• torchdynamo: None
• ray_scope: last
• ddp_timeout: 1800
• torch_compile: False
• torch_compile_backend: None
• torch_compile_mode: None
• dispatch_batches: None
• split_batches: None
• include_tokens_per_second: False
• include_num_input_tokens_seen: False
• neftune_noise_alpha: None
• optim_target_modules: None
• batch_eval_metrics: False
• eval_on_start: True
• use_liger_kernel: False
• eval_use_gather_object: False
• average_tokens_across_devices: False
• prompts: None
• batch_sampler: no_duplicates
• multi_dataset_batch_sampler: proportional

Training Logs

Click to expand

Epoch	Step	Training Loss
0	0	-
0.0091	100	0.7601
0.0183	200	0.4791
0.0274	300	0.1641
0.0366	400	0.0654
0.0457	500	0.0514
0.0549	600	0.0365
0.0640	700	0.0483
0.0732	800	0.0221
0.0823	900	0.0202
0.0915	1000	0.029
0.1006	1100	0.0215
0.1098	1200	0.0377
0.1189	1300	0.0351
0.1281	1400	0.034
0.1372	1500	0.0385
0.1464	1600	0.019
0.1555	1700	0.0314
0.1647	1800	0.0272
0.1738	1900	0.0363
0.1830	2000	0.0161
0.1921	2100	0.0315
0.2013	2200	0.0156
0.2104	2300	0.0327
0.2196	2400	0.0447
0.2287	2500	0.0251
0.2379	2600	0.0179
0.2470	2700	0.0185
0.2562	2800	0.0239
0.2653	2900	0.0268
0.2745	3000	0.0289
0.2836	3100	0.0312
0.2928	3200	0.0177
0.3019	3300	0.0283
0.3111	3400	0.0295
0.3202	3500	0.0335
0.3294	3600	0.0276
0.3385	3700	0.0232
0.3477	3800	0.0257
0.3568	3900	0.0164
0.3660	4000	0.0168
0.3751	4100	0.014
0.3843	4200	0.024
0.3934	4300	0.0169
0.4026	4400	0.0327
0.4117	4500	0.0269
0.4209	4600	0.0218
0.4300	4700	0.0399
0.4392	4800	0.0204
0.4483	4900	0.0183
0.4575	5000	0.0248
0.4666	5100	0.0171
0.4758	5200	0.0144
0.4849	5300	0.0255
0.4941	5400	0.0297
0.5032	5500	0.0186
0.5124	5600	0.0277
0.5215	5700	0.0187
0.5306	5800	0.028
0.5398	5900	0.0246
0.5489	6000	0.021
0.5581	6100	0.0186
0.5672	6200	0.0312
0.5764	6300	0.024
0.5855	6400	0.0273
0.5947	6500	0.0282
0.6038	6600	0.0177
0.6130	6700	0.012
0.6221	6800	0.0183
0.6313	6900	0.0186
0.6404	7000	0.0151
0.6496	7100	0.0233
0.6587	7200	0.0235
0.6679	7300	0.0249
0.6770	7400	0.0209
0.6862	7500	0.0195
0.6953	7600	0.0213
0.7045	7700	0.0298
0.7136	7800	0.0199
0.7228	7900	0.0183
0.7319	8000	0.0186
0.7411	8100	0.02
0.7502	8200	0.0232
0.7594	8300	0.0197
0.7685	8400	0.034
0.7777	8500	0.0153
0.7868	8600	0.0262
0.7960	8700	0.0218
0.8051	8800	0.0308
0.8143	8900	0.032
0.8234	9000	0.0131
0.8326	9100	0.018
0.8417	9200	0.0264
0.8509	9300	0.0208
0.8600	9400	0.0163
0.8692	9500	0.0158
0.8783	9600	0.0321
0.8875	9700	0.0238
0.8966	9800	0.0192
0.9058	9900	0.0148
0.9149	10000	0.0324
0.9241	10100	0.0254
0.9332	10200	0.0229
0.9424	10300	0.0132
0.9515	10400	0.0226
0.9607	10500	0.0213
0.9698	10600	0.022
0.9790	10700	0.0276
0.9881	10800	0.0312
0.9973	10900	0.0115

Framework Versions

• Python: 3.10.12
• Sentence Transformers: 3.3.1
• Transformers: 4.47.0
• PyTorch: 2.5.1+cu121
• Accelerate: 1.2.1
• Datasets: 3.2.0
• Tokenizers: 0.21.0

Citation

BibTeX

Sentence Transformers

@inproceedings{reimers-2019-sentence-bert,
    title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
    author = "Reimers, Nils and Gurevych, Iryna",
    booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
    month = "11",
    year = "2019",
    publisher = "Association for Computational Linguistics",
    url = "https://arxiv.org/abs/1908.10084",
}

MultipleNegativesRankingLoss

@misc{henderson2017efficient,
    title={Efficient Natural Language Response Suggestion for Smart Reply},
    author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
    year={2017},
    eprint={1705.00652},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}